Abstract
Based on a formal characterization of time-series and state-sequences, this paper proposes a new algorithm named the Optimal Temporal Common Subsequence (OTCS) to measure the similarity between state-sequences. Distinguishing from the conventional Longest Common Subsequence based measurements, a new concept of common subsequence named 'temporal common subsequence' is proposed to describe the similarity of the temporal order over state-sequences, as well as the similarity of the other two essential and vital temporal characters, i.e., the temporal duration of each state and the temporal gaps between each pair of adjacent states. The experimental results on news video retrieval demonstrate the effectiveness and validity of OTCS.
Original language | English |
---|---|
Title of host publication | 2nd International Conference on Software Engineering and Data Mining (SEDM) |
Publisher | IEEE |
Pages | 316-321 |
Number of pages | 6 |
ISBN (Print) | 9781424473243 |
Publication status | Published - 23 Jun 2010 |
Event | 2nd International Conference on Software Engineering and Data Mining (SEDM) - Chengdu, 23-25 June, 2010 Duration: 23 Jun 2010 → … |
Conference
Conference | 2nd International Conference on Software Engineering and Data Mining (SEDM) |
---|---|
Period | 23/06/10 → … |